Novel species of Cladosporium from environmental sources in Spain

Abstract Cladosporium is a monophyletic genus in Cladosporiaceae (Cladosporiales, Dothideomycetes) whose species are mainly found as saprobes and endophytes, but it also includes fungi pathogenic for plants, animals and human. Species identification is currently based on three genetic markers, viz., the internal transcribed spacer regions (ITS) of the rDNA, and partial fragments of actin (act) and the translation elongation factor 1-α (tef1) genes. Using this phylogenetic approach and from morphological differences, we have recognized six new species originating from soil, herbivore dung and plant material collected at different Spanish locations. They are proposed as Cladosporium caprifimosum, C. coprophilum, C. fuscoviride and C. lentulum belonging in the C. cladosporioides species complex, and C. pseudotenellum and C. submersum belonging in the C. herbarum species complex. This study revealed that herbivore dung represented a reservoir of novel lineages in the genus Cladosporium.


Introduction
Cladosporium is a ubiquitous genus in the family Cladosporiaceae of the recently proposed order Cladosporiales in the Dothideomycetes (Abdollahzadeh et al. 2020). Their species inhabit a wide range of substrates and have been reported to be among the most common fungi in both indoor and outdoor environments, including in extreme ecological niches (Flannigan et al. 2002;Bensch et al. 2010Bensch et al. , 2012Bensch et al. , 2018Sandoval-Denis et al. 2015;Temperini et al. 2018;Chung et al. 2019). Most Cladosporium species are saprobic, but some have also been reported as endophytes, hyperparasites on other fungi and plant as well as animal pathogens, including humans (Heuchert et al. 2005;Sandoval-Denis et al. 2016;de Hoog et al. 2017;Marin-Felix et al. 2017). Certain species show the ability to produce compounds of medical interest or are relevant as potential biocontrol agents for plant disease (Köhl et al. 2015;Khan et al. 2016;Adorisio et al. 2019).
Cladosporium is morphologically characterized mainly by its asexual morph, which shows differentiated conidiophores producing acropetal chains of conidia from monoor polyblastic conidiogenous cells. Both conidiogenous cells and conidia exhibit conidiogenous loci (scars) with a unique coronate structure, which is composed of a central convex dome surrounded by a raised periclinal rim, usually thickened, refractive and darkly pigmented (David 1997). Based on these features and DNA phylogeny derived from the LSU nrRNA gene, the genus has been well-delineated and distinguished from other cladosporium-like genera such as Hyalodendriella, Ochrocladosporium, Rachicladosporium, Rhizocladosporium, Toxicocladosporium, Verrucocladosporium and the recently described genus Neocladosporium Bezerra et al. 2017). Phylogenetic relationships among species of Cladosporium s. str. have been studied extensively over the last decade by a multi-locus sequence analysis approach with sequences of the internal transcribed spacers (ITS) of the rDNA and of the two protein encoding genes, translation elongation factor 1-α (tef1) and actin (act). The molecular approach combined with morphological features have allowed recognition of more than 230 species within the genus, which are split into three species complexes, i.e., the Cladosporium cladosporioides, Cladosporium herbarum and Cladosporium sphaerospermum complex Bensch et al. 2010Bensch et al. , 2012Bensch et al. , 2015Bensch et al. , 2018Sandoval-Denis et al. 2016;Marin-Felix et al. 2017).
While aiming to explore the diversity of microfungi from Spain, several interesting Cladosporium isolates have been recovered from different environmental samples. Using the above mentioned polyphasic approach and following the Genealogical Phylogenetic Species Recognition (GCPSR) criterion (Taylor et al. 2000), the taxonomy of those isolates has been resolved in six novel species for science; four pertaining to the C. cladosporioides species complex and two to the C. herbarum complex.

Samples and isolates
Samples of soil, plant debris and herbivore dung were collected between 2016 and 2018 at various Spanish locations. Dilution plating methods were used for isolating fungi from soil and dung samples following the procedure described by Crous et al. (2009) and a modified protocol described by Waksman (1922), respectively. In addition, soil samples were also processed by a baiting technique using small pieces of wood and filter paper as baits on the soil surface (Calduch et al. 2004). Samples of plant debris and also part of the herbivore dung were incubated in moist chambers following the procedures described by Castañeda-Ruiz et al. (2016) and Richardson (2001), respectively.
Among the cladosporium-like fungi found, we recovered eight isolates in pure culture which did not match any of the currently accepted species within the genus Cladosporium (Table 1). The isolates were deposited in the culture collection of the Universitat Rovira i Virgili (FMR, Reus, Spain) and, once phylogenetically and morphologically characterized, living cultures of the novel species and dry cultures for holotypes were also deposited in the Westerdijk Fungal Biodiversity Institute (CBS; Utrecht, the Netherlands). Nomenclatural novelties and descriptions were deposited in MycoBank (Crous et al. 2004).

DNA extraction, amplification and sequencing
Genomic DNA was extracted from cultures growing on potato dextrose agar (PDA; Pronadisa, Spain) after 7 days of incubation at 25 °C, following the modified protocol of Müller et al. (1998). Protocols listed previously in Sandoval-Denis et al. (2016) were used for amplification and sequencing. The primer pairs used were ITS5/ITS4 (White et al. 1990) to amplify the ITS region including the 5.8S gene of the rDNA, EF-728F/ EF-986R to amplify a partial fragment of the tef1 gene, and ACT-512F/ACT-783R to amplify a partial fragment of act gene (Carbone and Kohn 1999). PCR products were purified and stored at -20 °C until sequencing. The sequences were obtained using the same primers at Macrogen Europe (Macrogen Inc. Amsterdam, The Netherlands). Finally, the software SeqMan v. 7.0.0 (DNAStarLasergene, Madison, WI, USA) was used to assemble, edit and obtain the consensus sequences, which were then deposited in GenBank of the National Center for Biotechnology Information (NCBI) ( Table 1).

Sequence alignment and phylogenetic analysis
The sequences obtained were compared with other fungal sequences deposited in the NCBI database through the BLASTn tool. Alignment of those sequences and the phylogenetic analysis for each locus were performed with the MEGA (Molecular Evolutionary Genetics Analysis) program v. 6.0. (Tamura et al. 2013), using ClustalW algorithm (Thompson et al. 1994) and refined with MUSCLE (Edgar 2004) or manually if necessary, on the same platform. Since the isolates under study were related to the C. cladosporioides and C. herbarum species complexes, we also carried out alignments including sequence data of ex-type and reference strains of all the species from both complexes retrieved from the GenBank and mainly published by Schubert et al. (2007), Bensch et al. (2010, Sandoval-Denis et al. (2016) and Marin-Felix et al. (2017) (Suppl. material 1: Table S1). Phylogenetic reconstructions were made with the three phylogenetic markers (ITS, act and tef1) recommended for an accurate identification at the species level (Bensch et al. 2010(Bensch et al. , 2018Marin-Felix et al. 2017) using Maximum Likelihood (ML), Maximum Parsimony (MP), and Bayesian Inference (BI) analyses, with the Mega software v. 6.0. for the former two (Tamura et al. 2013) and with MrBayes v.3.2.6 for the latter one (Ronquist et al. 2012). Phylogenetic concordance of the three-locus datasets was evaluated through Incongruence Length Difference (ILD) implemented in the Winclada program (Farris et al. 1994) and also by visual comparison of the individual phylogenies in order to assess any incongruent results between nodes with high statistical support.
Determined by Mega software v. 6.0., the best nucleotide substitution model for ML analysis of the C. cladosporioides complex was General Time Reversible with Gamma distribution and invariant sites (GTR+G+I), and for the C. herbarum complex the best was the Kimura 2-parameter with Gamma distribution and invariant sites (K2+G+I). Bootstrap support value (MLBS) ≥ 70% was considered significant (Hillis and Bull 1993).
The MP analysis was performed using the heuristic search option with TBR (tree bisection and reconnection) branch swapping and 1,000 random sequence additions. Tree length (TL), consistency index (CI), retention index (RI), rescaled consistency index (RCI) were calculated. Bootstrap analysis was based on 1,000 replications. Maximum parsimony bootstrap support value (PBS) ≥ 70% was considered significant (Hillis and Bull 1993).
Determined by jModelTest (Posada 2008), the best nucleotide substitution models for the BI of the C. cladosporioides complex were Jukes Cantor with invariant sites (JC+I) for ITS, General Time Reversible with Gamma distribution (GTR+G) for tef1 and Hasegawa-Kishino-Yano with Gamma distribution (HKY+G) for act; and for the C. herbarum complex the best were the Kimura 2-parameter with Gamma distribution (K80+G) for ITS, Hasegawa-Kishino-Yano with Gamma distribution (HKY+G) for tef1 and act. The parameter settings used in these analyses were two simultaneous runs of 10,000,000 generations, and four Markov chains, sampled every 1,000 generations. The 50% majority rule consensus tree and posterior probability values (PP) were calculated after discarding the first 25% of the samples. A PP value of ≥ 0.95 was considered significant (Hespanhol et al. 2019).
Final sequence alignments and trees generated in this study were registered in Tree-BASE under the submission number S27350 (http://treebase.org).

Phenotypic studies
Microscopic features of the Cladosporium isolates were obtained from cultures growing on synthetic nutrient-poor agar (SNA; 1 g of KH 2 PO 4 , 1 g of KNO 3 , 0.5 g of MgSO 4 × 7H 2 O, 0.5 g of KCl, 0.2 g of glucose, 0.2 g of sucrose, 14 g of bacteriological agar, 1 L of distilled water) after 7 to 14 days at 25 °C in the dark, mounted onto semipermanent slides with Shear's solution (Bensch et al. 2018). At least 30 measurements were taken to calculate length and width ranges of the conidia and ramoconidia, given as the mean ± standard deviation in the descriptions. Macroscopic characterization of the colonies was made on PDA (Pronadisa, Spain), oatmeal agar (OA; 30 g of oatmeal, 13 g of bacteriological agar, 1 L distilled water) and SNA after 14 days of incubation at 25 °C in darkness. Colour notation of the colonies in descriptions were from Kornerup and Wanscher (1978). In addition, cardinal temperatures for the fungal growth were determined on PDA cultures after 14 days at temperatures ranging from 5 to 40 °C at intervals of 5 °C.

Phylogeny
Three individual phylogenies (ITS, tef1 and act), carried out for the C. cladosporioides and C. herbarum species complexes, were visually very similar and the ILD test showed that the three loci datasets were congruent in both complexes (P = 0.16) and could be combined. Phylogenies obtained by ML, MP and BI also showed a visual topological congruence and were similar to that obtained by other authors (Marin-Felix et al. 2017;Bensch et al. 2018). The combined alignment of the three mentioned loci datasets encompassed 101 sequences in the C. cladosporioides complex and 58 sequences in C. herbarum complex. The alignment for the former group comprised 1,060 bp (ITS 484 bp, tef1 313 bp and act 263 bp), which included 424 bp variable sites (ITS 47 bp, tef1 239 bp and act 138 bp) and 319 bp phylogenetically informative sites (ITS 25 bp, tef1 193 bp and act 101 bp). Two species of the C. sphaerospermum complex, C. sphaerospermum CBS 193.54 and C. longissimum CBS 300.96, were included as outgroup in this first multi-locus phylogeny (Fig. 1). For the maximum parsimony analysis the maximum of 1,000 equally most parsimonious trees were saved (Tree length = 1614; CI = 0.294; RI = 0.666; RCI = 0.214).
The eight unidentified isolates did not match any known lineage of Cladosporium species, six were related to the C. cladosporioides species complex and two to the C. herbarum complex, and together they represented six new phylogenetic species in the genus.
In the combined phylogeny of the C. cladosporioides complex, 71 species were delineated (Fig. 1). The isolates FMR 16101 and FMR 16164 formed a strongly supported terminal clade representative for a unique taxon, but with an uncertain phylogenetic position due to the low statistical support (-MLBS / 77 PBS / -PP) for the nearest lineages of C. chasmanthicola and C. sinuatum. Both unidentified isolates were genetically identical and showed a percentage of identity with the ex-type strains of these latter species of 97.22% and 97.65% for act, and 96.79% and 97.50% for tef1, respectively. A second undescribed monophyletic terminal clade included FMR 16288 and FMR 16389, which grouped with the lineages of C. exasperatum, C. parapenidielloides and C. longicatenatum in a clade with highly supported values (94 MLBS / 84 PBS / 0.98 PP). However, both isolates showed a sufficient genetic distance to be considered a distinct species from the closest, C. longicatenatum and , with a sequence similarity of 95.75% and 95.28% for act and 90.87% and 90.48% for tef1 respect to the ex-type strains of these two known species. FMR 16532 and FMR 16385 formed two distinct monophyletic branches. The former showed an uncertain phylogenetic position with the species in the complex; the comparison of its sequences with those of the GenBank dataset through the BLASTn tool showed that the ITS was 100% similar with several species of the C. cladosporioides complex, while sequences of act and tef1 were 99.12% and 89.02% similar with sequences belonging to C. asperulatum (UTHSC DI-13-216/GenBank LN834541 and CBS 113744/GenBank HM148237, respectively). FMR 16385 was closely related to the ex-type strain of C. alboflavescens (100 MLBS / 100 PBS / 1 PP). The percentages of identity between these latter two fungi (97.79% for act and 96.75% for tef1) together with morphological differences observed allow us to consider them distinct taxa.
In the C. herbarum complex, 40 species were phylogenetically well-delimited, including two novel lineages each represented by FMR 16231 and FMR 17264 (Fig. 2). Both were genetically and morphologically differentiated from their closest relatives, C. tenellum and C. subcinereum, respectively. The percentages of identity observed between the isolate FMR 16231 and the ex-type strain of C. tenellum (CBS 121634) were 97.78%, 83.76% and 100% for act, tef1 and ITS, respectively, and between FMR 17264 and the ex-type strain of C. subcinereum (CBS 140465) were 98.57%, 95.98% and 100% for act, tef1 and ITS, respectively.
The percentages of identity between the six putative new Cladosporium species and their relatives are summarized in Table 2. The novel taxa are described and illustrated in the taxonomy section below.
Notes. Although C. caprifimosum clearly belongs to the C. cladosporioides species complex, our multi-locus analysis did not reveal any phylogenetic relationship with other species in the complex. It is represented by a single branch placed distance from other Cladosporium species (Fig. 1). Cladosporium caprifimosum differs from the other novel species proposed here mainly by its aseptate and smooth conidia.  Etymology. Name refers to the substrate where the species was isolated, unidentified herbivore dung (ancient Greek, kópros = dung + phílos = loving). Type. Spain, Extremadura, Badajoz province, Granja de Torrehermosa, unidentified herbivore dung, Jan. 2017, J. Cano (holotype CBS H-24470; cultures ex-type FMR 16164, CBS 144919).
Notes. Our phylogeny shows C. lentulum included in a well-supported terminal clade together with the ex-type strains of C. exasperatum, C. parapenidielloides and C. longicatenatum, three species all described from plant material collected in Australia (Bensch et al. 2010(Bensch et al. , 2015. However, the genetic distance allows it to be considered a distinct species within the clade (Fig. 1). Phenotypically, C. lentulum can be distinguished from its counterparts mainly by its slower growth, especially on OA at 25 °C after 14 d (19-20 mm vs 39-54 mm for C. exasperatum, 42-55 mm for C. parapenidielloides and 43-54 mm for C. longicatenatum). In addition, our new species shows shorter ramoconidia (10.5-23 μm) than C. exasperatum and C. longicatenatum (19-40 μm and 22-42 μm, respectively); ramoconidia in C. parapenidielloides are absent; the conidia in C. lentulum are smooth or nearly so, while those of C. exasperatum and C. longicatenatum possess a unique verruculose-rugose conidial surface ornamentation, especially prominent in the former; and conidiophores in C. parapenidielloides are much shorter (up to 67 μm) than those observed in C. lentulum (up to 406 μm) (Bensch et al. 2010(Bensch et al. , 2015.
Distribution. Spain. Notes. Based on the phylogeny of the C. herbarum complex (Fig. 2), C. pseudotenellum is closely related with C. tenellum, a species originally described from hypersaline water in Israel, later found on Phyllactinia sp. (Erysiphaceae), and in indoor air samples collected in the USA Bensch et al. 2012Bensch et al. , 2018. Our species differs from C. tenellum in the absence of micronematous conidiophores and in having shorter macronematous conidiophores (up to 146 μm vs up to 200 μm), shorter conidiogenous cells (15-32 μm vs 6-40 μm), with few conidiogenous loci (up to five vs up to 10 or more in C. tenellum), and shorter ramoconidia (9-14.5 vs up to 32 μm). In addition, terminal and intercalary conidia in C. pseudotenellum are aseptate, while those of C. tenellum are 0-1(-3)-septate Bensch et al. 2012).

Discussion
Cladosporium is a well-delineated genus, the taxonomic structure and phylogenetic relationships of its species have been investigated in several studies over the last decade, so far giving rise to a genus of more than two hundred well-established species Bensch et al. 2010Bensch et al. , 2012Bensch et al. , 2018Sandoval-Denis et al. 2016;Marin-Felix et al. 2017;Crous et al. 2009Crous et al. , 2019Jayasiri et al. 2019). However, this species number will continue to expand through the study of soil, which is a proven pool of fungal species that remains undescribed, and other substrates poorly investigated by molecular tools for fungal diversity (Tedersoo et al. 2017;Hyde et al. 2018). In this context, a set of Cladosporium isolates were obtained in pure culture from samples of soil, dung from different herbivorous animals, and plant debris collected during a survey of microfungi in various Spanish locations. Using the molecular approach for species delineation in Cladosporium (Bensch et al. 2012;Marin-Felix et al. 2017), eight of those isolates represented six novel lineages for the genus which are proposed as C. caprifimosum, C. coprophilum, C. fuscoviride, C. lentulum, C. pseudotenellum and C. submersum. Of note is that almost all the specimens in the present study (7/8) were isolated directly from the natural substratum incubated in moist chambers or from baiting technique plates. Although Cladosporium isolates are commonly detected by plating methods, the slow growth rate or the low spore concentration of some cladosporium-like fungi compared to other fungi present in a given substrate is probably a handicap to detection and/or isolation of uncommon Cladosporium species. Therefore, as recommended by Crous (1998) for similar fungi, techniques based on fungal isolation directly from the natural substratum should be considered a choice for future studies of Cladosporium species diversity.
To our knowledge, Cladosporium species as dung inhabiting fungi have been reported in a very few studies, C. cladosporioides and C. herbarum being the most reported species (Bell 1975;Seifert et al. 1983;Jeamjitt et al. 2006;Masunga et al. 2006;Piontelli et al. 2006;Simões-Calaça et al. 2014;Thilagam et al. 2015). However, in all those studies, fungal identification was based exclusively on morphological features. Only C. herbarum has been reported recently from crown droppings and identified molecularly, but using only the ITS barcode (Torbati et al. 2016). In our case, the three new species isolated on herbivore dung (i.e., C. caprifimosum, C. coprophilum, and C. lentulum) showed the typical morphological features attributed to the C. cladosporioides species complex. However, their identifications would have been difficult with morphological features alone, even with the analysis of their ITS sequences (Table 2) since they are identical under the universal barcode for fungi as reported in previous studies for many other Cladosporium species (Bensch et al. 2010(Bensch et al. , 2012Marin-Felix et al. 2017). Therefore, only sequence analysis with act and tef1 will allow us to know the real diversity of Cladosporium species from this understudied substrate by molecular tools.
Although no temperature studies have been systematically applied to characterize most Cladosporium species (Bensch et al. 2012(Bensch et al. , 2015(Bensch et al. , 2018, we agree with Ma et al. (2017) that cardinal temperatures for growth can help to differentiate certain species in their respective complexes. While species in the C. sphaerospermum complex show a maximum temperature for growth of no more than 30-32 °C, C. halotolerans was able to grow at 35 °C (Sandoval-Denis et al. 2015). Similarly, although most species of the C. cladosporioides complex do not tolerate high temperatures, C. angulosum, C. angustisporum, C. anthropophilum, C. flavovirens, C. funiculosum, C. pseudocladosporioides, C. subuliforme and C. tenuissimum were able to grow at 35 °C (Sandoval-Denis et al. 2015. To date, no member of the C. herbarum complex was found to be able to grow above 30 °C; however, one of the novel species of the complex described here, C. submersum, had a maximum growth at 35 °C. On the contrary, the recently described species C. neopsychrotolerans and C. tianshanense from the complex C. cladosporioides and C. psychrotolerans from the complex C. sphaerospermum showed a psychrophilic behavior Ma et al. 2017), demonstrating in part the ability of Cladosporium species to adapt to different environmental conditions.